7 rY'

May 13, 1970 - Neutral Carboranbs Related to BgHio n » 0° m . 1” m = 2°. m i 3° m = 4'. rY. 1. 'rY -1. 7 rY'. -1. ' rY. -1. ( ~\ ,. 6. 3 + w. Yr...
1 downloads 0 Views 572KB Size
Inorganic Chemistry, VoZ.10, No, I , 1971 181

AMINOACIDDERIVATIVES

I

a The irn,t y I , ,

15

rotated by 2lfm 5 belore beinq added to (he pentagonal m e unll.

pressed as a combination of the "equivalent" structures of columns 1 and 4. In C3B3H,,34the preferred structures are probably those in columns 2 and 4, which minimize the formal charges. The situation in the recently prepared C4BzHea5is less clear, but formal charge considerations suggest a preference for either the structures of column 0 or the equivalent pair, columns 1 and 4. The suggestion here of a ?r-donation model involving two double bonds would make a structural determination and comparison of C-C and C-B bond distances in this molecule extremely interesting. (34) C. L. Bramlett and R. N. Grimes, J. Amev. Chem. Soc., 88, 4269 (1966). (35) T.P.Onak and G. T. F. Wong, ibzd., in press; T. P. Onak, private communication, 1970.

The formalism developed above may also be applied to borane or carborane ions of the above species, such as BeH9- or Bd2zH7-.37 'An ion produced by removal of a proton from a bridge position is simply assigned those resonance structures of the parent isostructural carborane which minimize the formal charges. Acknowledgments.-We wish to thank the National Science Foundation for predoctoral fellowships to I. R. E., J , A. T., and E. S. and the Office of Naval Research for financial support of this work. (36) H.D. Johnson, 11, Ph.D. Thesis,The Ohio State IJniversity, 1969; Diss. Abslr., in press. (37) T. P. Onak and G. B. Dunks,Inoug. Chem., 6, 439 (1966).

CONTRIBUTION FROM THE DEPARTMENT OF CHEMISTRY, ST.LOUISUNIVERSITY, ST.LOUIS,MISSOURI 63103

Reaction of Amino Acid Derivatives with Trihaloboranes and with Sodium Borohydride BY E. F. ROTHGERY*

AND

L. F. HOHNSTEDT

Received May 13, 1970 The borane, trichloroborane, and some trifluoroborane adducts of several amino acid esters were prepared. The nmr and ir spectra show them all to be amine adducts. The BF3 adducts of methyl glycinate and of methyl palaninate were converted to the corresponding N-substituted borazines. Possible intramolecular coordination of ring borons by carbonyl groups in the N substituent was investigated but none could be observed. The reaction of BCl3 with glycine gave the novel compound B(O2CCHnNHs+)s,(BCla-)a.

Introduction The intramolecular coordination of a boron atom to give a four-coordinate species has been shown in many cases to result in a material that is often much more resistant to hydrolysis than is a similar three-coordinate compound. Diphenylborinic acid is very difficult to obtain, but its P-aminoethyl ester, I, has been isolated

* To whom correspondence should be addressed.

and found to be very resistant t o hydro1ysis.l This has C85, ,C&5 H,N**B,o

Hb--CH2 I

I

been attributed to intramolecular coordination of the (1) R.L. Letzinger and I. Skoog, J . Ameu. Chem. Soc., 77, 2491 (1955).

E. F. ROTHGERY AND L. F. HOHNSTEDT

182 Inorganic Chemistry, Vol. 10, No. 1, 1971 TABLE I ANALYTICAL DATA yocalcd

Mp, o c

CHaOC(O)CHzN(CH3)z ' BC13 CHaOC(O)CHzNHz*Be13 CHaOC(O)CHzI\;Hz' BF3 CH3OC(O)CHzCHiNHz BF3 CH30C(O)CHzNHz BH3 C H ~ O C ( O ) C H Z K ( C HBH3 ~)~. C~H;OC(O)CHZNH~CH~',BC~~(FBNCHzCHzC(O)OCHs)3 CHaCiX. B(OC(O)CHzKHa+,BC14--)3

92-94 53-55 90-9 1 Oil 70-71 24-26 Oil 42-45 95-100

5

C

H

B

Yo found------

-,--

7

Compound

C

N

CI

H

B

N

CI

25.82 4.69 4.61 45.40 5.98 25.75 5.03 4.87 44.73 5.98 17.46 3.39 5.24 51.45 6.79 17.83 3.96 4.70 51.21 7.08 5.09 6.92 9.94 24.55 10.02 25.57 5.01 7.67 8.20 29.66 6.41 5.81 28.71 5.31 6.32 8.86 3.61 35.81 9.11 10.14 35.01 9.79 10.50 3.33 10.70 43.16 10.31 8.29 45.85 10,77 8.25 10 60 5.17 21.35 4.59 4.54 52.00 5.17 22.88 4.47 3.99 52.37 10.70 37.37 6.05 8.45 36.70 5.39 8.26 9.87 5.16 48.92 5.28 51.96

boron by nitrogen giving a five-membered ring. Carbonyl groups may also act as ligands and coordinate boron.*r3 When dialkylchloroboranes (RtBC1) are treated with ethyl acetoacetate, they give compounds of unusual hydrolytic and oxidative stability, 11. Stability has been attributed to ring formation and resonance.

I

borohydride and diborane are capable of reducing esters and trichloroborane cleaves ester groups.7 If the borazines are obtained and intramolecular coordination occurs, the two primary amino esters studied would result in five- and six-membered side rings, the most favorable size for ring systems. Experimental Section

Materials.-Methyl glycinate hydrochloride, methyl 8-alaninate hydrochloride, and ethyl sarcosinate hydrochloride were C,H,O-C ACH, prepared from the respective amino acids via the Fischer esterifiC' ' H cation, using excess hydrogen chloride. Since the neutral amino esters rapidly peptize, they had to be I1 prepared immediately before using. This was done by bubbling ammonia through a diethyl ether slurry of the ammo ester Also of interest is the reaction of trichloroborane with hydrochloride a t O " , filtering the XHaC1, and vacuum distilling carboxylic acids. If acetic acid is treated with BC& in CH302CCHzNH2, chloroform, the product obtained is (ACO)~BOB (OAC)~, the ester.* Boiling points are as follows: HZ, bp 31' (6 mm), lit.8 bp 45' (20 mm); C H ~ O Z C C H ~ C H ~ Nhp in which the borons are coordinated by carbonyl oxy31' (5 mm), lit.9 bp 58' (15 mm); CH3CH20zCCH2XHCH3; bp gens4S6 If the same reaction is carried out in dioxane, 48" (10 mm), 1it.lO bp 60" (25 mni). Methyl N,N-dimethylglycinate ufasprepared by the reaction of the product obtained is d i o x . B ( O A ~ ) ~In . ~this species methyl chloroacetate nith dimethylamine, bp 36" (8 mm), lit.@ the fourth position on the boron is taken by the solvent. bp 51" (12 mm). We decided to attempt the preparation of some borSpectra.-Kmr spectra were obtained on a Varian HA-100 azines which contained carbonyl groups in the nitrogen spectronieter, the proton spectra a t 100 MHz and boron a t substituent and to determine whether or not the oxygen 32.1 MHz. Infrared spectra were recorded on a Perkin-Elmer Model 457 or a Beckman IR-5A spectrophotometer. would coordinate the ring borons and thus have any Trichloroborane Adducts.-Samples of the amino esters on effect on the hydrolysis of the borazine. the order of 1 g were dissolved in CH2C12 or CHC13 and

n

r

Because they had not been previously reported, we first prepared the borane and some trihaloborane adducts of several amino esters in order to establish the donor atom. This was required because trihaloboranes are capable of forming adducts with carbonyl groups as well as with amines. Esters were employed because the use of amino acids would lead to borates. The methyl esters of glycine, @-alanine,and N,N-dimethylglycine and the ethyl ester of N-methylglycine (sarcosine) were used. The boranes formed amine adducts in all cases, including the two primary amino esters, methyl glycinate and methyl p-alaninate, thus allowing the possibility for their conversion to the corresponding N-substituted borazines by various routes such as pyrolysis and addition of triethylamine, etc. The standard preparative methods for borazines are somewhat limited when using amino esters in place of alkyl- or arylamines. Sodium (2) W. Gerrard, Yi.F. Lappert, and R. Shepperton, J . Chem. Soc., 3648 ( 1958).

(3) E. Muetterties, "The Chemistry of Boron and I t s Compounds," Wiley, New York, N. Y . ,1967. J . Chem. Soc., 3652 (1968). (4) L. A . Duncanson, et d., ( 5 ) W. Gerrard and M . A. Wheelans, Chem. I n d . (London),758 (1954). ( 6 ) H. V. Ribbel, Z . Anorg. Allg. Chem., 369,203 (1968).

TABLE I1 1H NMROF TRIHALOBORANE ADDUCTS Compound

Group

Chemical shift, ppm

Coupling, Hz

0

I1

C H ~ O C C H Z S ( C HBe13 ~)~. (CHC13)

3.25 4.24 3.82

q,2.8 b s

3.61 3.79 5.15

t, 4 s b

0

,I

CH30CCHzNHz. BF3 ( CHa C S ) 0

I1

CHaOCCHzCHehTHz* BF3 (CHz Clz)

3.69 2.69 3.16 5.87

s t, 6 . 3

t, 6 . 3 b

_____ (7) M. Frazer and W. Gerrard, J. Chem. SOC., 2959 (1955). ( 8 ) M . Franckel and E . Kalchaliki, J . Amer Chem. Soc., 64, 266 (1942). (9) B. J. Heilbron, Ed., "Directory of Organic Compounds," Oxford Press, Kew York, N. Y.,1946. (10) E, A. Pritt and S, M. McElvain, J . Amer. Chem. Soc., 66, 1233

(1933).

Ivorganic Chemistry, Vol. 10, No. 1, 1971 183

AMINOACIDDERIVATIVES

TABLE V lH NMROF BORAZINES

TABLE I11 'H NMROF BORANE ADDUCTS Compound

Chemical shift, ppm

Group

Coupling,

Hz

Compound

Chemical shift, ppm

Group

Coupling, Hz

0

0

/I

I/

CH30CCHgNHz BH3 (CHzClz)

-CHz-OCA??, -NHz -BH3

3.61 3.79 4.46 Not obsd

(FBNCHzCOCHa)3 (CHzClz)

t, 7 S

b

/I

(FBNCHzCHzC0CHa)s (CHzCL)

CH3CHzOCCHzNHCHs.BH3 -CHzCHs ( CHzClz) -NCHa -NCHz-CHtCHs -NH -BH3

1.29 2.59 -3.5 4.25 4.90 Not obsd

t, 7.2 d, 6 AB q, 7.2 b

( CHzClz )

s

-NCHzCHz -NCHzCHz-0CH3

2.47 3.40 3.64

t, 7 t, 7 s

Compound

-N(CH3)z -NCHz-0CHs -BH3

2.83 3.66 3.76 Not obsd

I1

S

S

CHaOCCHzNH3+, BCla(CHaCN) 0

S

TABLE IV CHsCHzOCCHzNHzCHa+,BClrAMINOESTERHYDROTETRACHLOROBORATES (CHzClz) Group

Chemical shift vs. (C2Hs)zO. BF3, Coupling, ppm HZ

0

ll

1H NMROF

s

TABLE VI llB NMRSPECTRA

0

II CH30CCHzN(CHa)z*BH3

3.71 3.89

0

0

II

-0CH3 -CHz

Chemical shift, ppm

S

-6.5

S

-6.1 -5.0

S

0

coupling,

Hz

-6.6

/I

CHaCN. B(OCCHzNHa+,BC14-)3 (CHsCN)

S

0

-0CHs -CHz-NH3+

3.81 3.82 6.64

s q,5.8 b

-CHzCH3 -NCH3 -NCHz-CHzCHa -NHz+

1.28 2.73 3.83 4 27 7.35

t,,7.O t, 7 . 5 t, 7 . 5 t, 7 . 0 b

-CHz-NH3 +

3.78 6.51

q,5.3 b

-CHz-NH3+

3.74 6.50

q,5.3 b

I

0

II

B(OCCHzNHa+,BC14-)3 (CHsCN) (from glycine) 0

/I

B (OCCHzNHaf,BC14-)a (CH3CN) (from glycine hydrochloride)

was condensed onto the frozen solutions a t -196'. CH30C(O)CH2N(CH3)z.BCl3 gave a yellow solution. Removal of the solvent and extraction of the solids with benzene gave a pure product. C Z H ~ O C ( O ) C H Z K H C H ~ . Brapidly C~~ disproportionated. Spectra were obtained, but no sample for analysis could be isolated. The adducts may also be prepared using the solid adduct CH&NsBC13 instead of gaseous BC13. Trifluoroborane Adducts .--Freshly distilled amino esters were dissolved in diethyl ether a t -20' under an Nz cover and (CZH~)ZO.B was F ~ slowly added causing the formation of a white oil. The ether was decanted and the product was reprecipitated from CHIC12 to remove any (C$He)20.BF3. Pumping on the oils sometimes caused them to change to solids but usually the solids reverted t o oils on standing. Borane Adducts.-The amino ester hydrochloride and NaBH4 were refluxed overnight in monoglyme under an Nz cover. A slight excess of NaBH4 was used. The hot reaction solution was filtered under Nz. Removal of the solvent left a white solid.

11

CH30CCHzNHz. BH3 (CHzClz) 0

ll

CHsCHzOCCHzNHCHa * BH3 (CHzClz) 0

$17.5

q, 96

+13.6

q, 96

$6.8

q, 96

II

CHaOCCHzN( CH3)z * BH3 (CHzC12) 0

I/

(FBNCHzCHzCOCH3)a (CHzClz)

-35.7

S

C Z H ~ O C ( O ) C H Z N H C H ~ *was B H ~prepared in this manner, but no sample suitable for analysis was obtained. CHsOC(0)CHzK(CH3)z.BHa was prepared by adding BzH6 to a chloroform solution of the amino ester a t - 196". All the amino ester adducts showed uncoordinated carbonyl stretch a t 1720-1740 cm-1. The borane adducts all showed B-H stretches a t 2250-2350 cm-'. Tetrach1oroborates.-Acetonitrile-trichloroborane (0.9160 g, 5.789 mmol) and methyl glycinate hydrochloride (0.7282 g, 5.793 mmol) were placed in a flask with a stirring bar and put on a vacuum line, and the flask was evacuated. Thirty milliliters of CH&N was condensed onto the solids a t -196". On warming to room temperature all the solids dissolved. Removal and fractionation of the solvent showed xi0 HC1 to have been evolved. The solid product left after removal of solvent melted a t 53-55', but accurate analysis could not be obtained. Since the BClasalt is soluble in CHsCN, in which CH30C(O)CHzNH3+,C1- is insoluble, larger amounts of the product were prepared by treating excess amino ester hydrochloride with CH3CN*BClsunder Nz, filtering, and recovering the product from the filtrate. CZH~OC(O)CHZNHZCH~+,BC~~was prepared in CHzClz from the corresponding hydrochloride salt and Bel3. The BC14-salt is somewhat soluble in CHzClz.

184 Inorganic Chemistry, Vol. 10, No. 1, 1971 Both products reported showed strong absorptions in the ir spectrtim a t 665 and 690 cm-' (BCl4-). CH3CN.B (OC( O)CHzNH3+,BC14-)3.-Acetonitrile-trichloroborane (0.3324 g, 2.108 mmol) and glycine (0.1183 g, 1.576 mmol) were mixed in a 50-ml flask. The flask was placed on a vacuum line and evacuated, and 25 rnl of CH3CN was condensed in a t -196'. On warming t o room temperature all the solids dissolved. Fractionation of the solvent through traps a t -78, -126, and -196" showed no HC1. -4 reaction was performed as above using glycine hydrochloride (0.2754 g, 2.47 xnmol) and CHxCN .BC13 (0.5208 g, 3.29 mmol). Fractionation of the solvent as above gave 2.36 mmol of a gas identified as IIC1. The products obtained from glycine and from glycine hydrochloride were shown to be identical by their melting point (95-100°), ir [1720 (C=O), 665, and 690 cm-I (BCL-)], llB and 'H nmr, and analysis. Although analytical data are not in close agreement with the proposed adduct, the B:C1 ratio is that expected for the proposed tetrachloroborate. Trifluorob~razine.-(FBNCH~CH,CO~CH~)~ was prepared by mixing C H ~ O C ( O ) C I I ~ C H ~ N(8.364 HZ g, 0.08 mol) and (iCaH,)dYC2H: (20.756 g, 0.161 mol) in benzene and slowly dropping in a benzene solution of (C2Hs)20.BF3 (34.191 g, 9.241 mol).ll On the addition heat was evolved, a solid formed, and the solution turned yellow. The mixture was stirred a t room temperature for 15 hr. Filtration and removal of the solvent left a light brown oil. The oil was reprecipitated from CH2C12 with hexane. On one occasion a solid formed which later reverted to an oil. I r spectrum (film): 1730 (C=O), 1490-1450 (BY), 1200-1170 (BF), and 720 cm-' (BN); strong absorption also a t 720 an-'. Mass spectrum (70 ell) m/e (relative intensity): 374 ( l ) , 320 (l), 91 (59), 92 (46), 73 (41), 61 (42), 43 (55), 44 ( 5 5 ) , 29 (74), 30 (94), 31 (loo), 15 (63). (FBNCH2C02CH3)3was prepared as above using freshly prepared methyl glycinate. Pure samples of product could not be obtained. Ir spectrum (film): 1730 (C=O), 1490-1440 (BPU'), 1200-1150 (BF), and 760 cm-' (BX). Mass spectrum (70 el') m / e (relative intensity): 352 (l), 292 (61), 175 (39), 91 (65), 84 (74), 73 (65), 61 (61), 59 ( 6 l ) , 49 (loo), 30 (83), 15 (60). The fragments reported are those of large relative intensity, those matching in the spectra of the two borazines or those attributable to ions expected from a borazine with an ester side group. Each of the mass spectra contains several hundred peaks, presumably due to thermal decomposition as well as fragmentation. I t did not appear valid to assign peaks on the basis of differences in boron isotope content. Attempted Preparation of (ClBNCH2C02CH3)3and of (HBNCH2CO~CH3)3.--Numerous attempts were made to prepare the N-substituted B-trichloroborazine from CH30C(O)CHp KH2.BCl3 and from CH30C(O)CHs~H3+,BCla- by heating and/or addition of (C2H5)3Xin a variety of solvents. However, in each case only ( C ~ H S ) ~ I \ ; H + and/or C~a black, tarry material resulted. As might have been expected, refluxing CH30C(O)CHzr\"3+, C1- with NaBHa in dislyrne apparently reduced the ester. White Xmr specneedles formed in the relux condenser, mp 43-46'. trum (CHCl3): 6 3.45 (s, ?). Ir spectruni (CHCL): 1475 ( B N ? ) and 1335 cm-l (BO?); no (C=O) or (BH) frequencies. No desired product could be obtained by pyrolysis of cI-130c(O)CH2XH2 .BH3, either with or without solvenrs.

Discussion Adducts.-The adducts of (CH3)2NCH2C02CH3 were prepared first in order to establish the donor site. A tertiary amino ester was chosen since they are the least moisture sensitive and are not prone to decomposition. The coupling of IlB (spin 3/2) with the N(CH& group giving rise to a 1: 1: 1: 1 quartet for the methyl proton shows nitrogen unquestionably to be the donor atom. This coupling occurs only with BC13. Neither the (11) J. J. Harris and B. Rudner, Inovg. C h e m . , 8 , 1258 (1969).

E. F. ROTHGERY AND L. F. HOHNSTEDT BH3 nor BF3 adduct shows coupling of boron with the N-methyl protons. This same pattern of coupling has been observed for the same three borane adducts of trimethylamine.12 All three adducts exhibit considerable deshielding of the N-methyl and N-methylene protons in the order BCls > BF3 BH3. The methoxy protons being six bonds removed are only slightly affected by adduct formation. The ir spectra confirm that no carbonyl adduct is formed, since there is no change in carbonyl stretching frequency. Ethyl acetate shows a shift of 170 cm-' in its carbonyl stretch on addition of BC13.13 The secondary amino ester adducts are by far the least stable. C13B. C H ~ H N C H ~ C O ~ Ccould ~ H Z never be isolated, probably because of rapid disproportionation to an aminoborane and a tetrachloroborate. Absorptions were noted in the ir arising from a BC14species. The borane adduct, H3B. CH3HNCH2C02C2H6,was isolated, but it slowly evolved a gas, probably hydrogen. The amine-borane section of H3B. CH3HNCH2C02C2H6exhibits an interesting ninr spectrum (Figure 1). The CH3 group appears as a doublet because

-

5!0

Figure 1.-The

4!0

r!o

lH nmr spectrum of ethyl sarcosinate-borane.

of coupling with the amine proton. No splitting by the boron is observed. The N-CH2- group is adjacent to an asymmetric nitrogen and as a result the two hydrogens are nonequivalent, coupling with one another as well as with the amine proton giving the eight peaks of the AB portion of an ABX spectrum, On inspection of the spectrum of the BC1, adduct o€ ethyl sarcosinate in which IlB coupling is observed a very interesting phenomenon is noted (Figure 2 ) . The N-CH3 group appears as a septet. This arises from coupling with boron (spin 3 / 2 ) , J B = ~ 2.8 Hz, and with the amine proton, J = 6.6 Hz. The two patterns overlap giving only six instead of eight peaks. The nitrogen atom having four different groups attached is again asymmetric and the two N-CH2 protons nonequivalent. The high-field half of the AB pattern consists of four peaks as does the same portion of the borane adduct (12) J. Miller and M D Onyschuk, C a n J Chem , 42, 1518 (1964). (13) M. F. Lappert, J. C h e m . Soc., 817 (1961).

Inorganic Chemistry, Vol. 10, No. 1,1971

AMINOACIDDERIVATIVES

C$.22"$

a!o

4!0

Figure 2.-The lH nrnr spectrum of ethyl sarcosinate-trichloroborane N-methylene and N-methyl groups.

spectrum. However the low-field half consists of two sextets as a result of coupling of one of the methylene hydrogens with boron as well as with the amine hydrogen. The primary amino ester adducts show a narrow solubility range; in fact no solvent was found for c13B.H~NCHZCOzCH3. It reacted violently with protic solvents. The BF3 adducts of methyl glycinate and of methyl p-alaninate and the borane adduct of methyl glycinate have some solubility in solvents usable for spectra. The spectra show them to be amine adducts; thus the path to possible conversion to the corresponding borazines is open. Another route to N-substituted-B-trichloroborazines is via the ammonium tetrachloroborates. In this work the use of amino ester hydrochlorides to form tetrachloroborates and then the borazines would have been convenient since the neutral amino esters are unstable and must be prepared each time a reaction is to be run. The inability to convert these salts to the trichloroborazines was probably caused by the reactivity of some boron chloride species toward the carbonyl groups. Ester groups are readily cleaved by trichloroborane. l 4 The reactions of glycine (eq 1) and of glycine hydrochloride (eq 2) with BC13 both give the same product. 4BCla

CHaCN + 3HsNCHrC02H -+ B(0zCCHzNHa)s3+ f

3BCla4BC13

+ 3C1-,HatNCH2C02H

(1)

CHaCN

-+B(0zCCHzNHs)s3+ f 3BClr-

+ 3HC1

(2)

Only with glycine hydrochloride is HC1 liberated. Varying the ratio of BC13 to glycine hydrochloride gave amounts of HC1 predicted by eq 2. The IlB nmr of acetonitrile solutions of the product showed two kinds of boron in approximately a 3: 1 ratio. The larger peak at -6.1 ppm is due to BCld-. The smaller peak a t -5.0 pprn must arise from the borate boron. Both signals are in the range of boron resonances in which tetrahedrally coordinated boron is found. An ir spectrum in KBr of the solid obtained on removing the (14) W. Gerrard and M. A. Wheelans, J . Chem. SOC.,4296 (1956).

185

solvent showed a C-N stretch a t 2350 cm-1 which indicates coordination by the nitrile group. The C-N stretch of pure acetonitrile is 2270 cm-1, while in CH,CN BC1, it is 2400 cm-l.I5 Trifluoroborazines.-The only borazines that were prepared are the two trifluoroborazines derived from methyl glycinate and methyl 0-alaninate. They were produced using the method recently described by Harris and Rudner.l' Their method is especially suitable to these systems for several reasons. Trifluoroborane is much less reactive toward carbonyl groups than is BCl3 and does not have the reducing properties of BH3. The ability to convert the adducts to the borazine by a method not requiring heating is favorable since it is less destructive of the amino esters. The trifluoroborazines were prepared a number of times and in each instance (FBNCHZCHZCOZCH~)~ was obtained in purer form than was (FBNCHZCOZCH3)3. On comparison of the nmr spectrum of F3B.H2NCH2CHZCOZCH, with that of (FBNCHZCHZCOZCH~)~ it is noted that in the borazine the triplet arising from the methylene group 01 to the nitrogen is deshielded by 0.24 pprn while the p-methylene is shielded by 0.22 ppm as compared to the adduct. This phenomenon is usually associated with groups substituted on aromatic rings. The mass spectra of the two trifluoroborazines were run only as an aid in identifying the compounds and no detailed mass spectral study was intended. The spectra were run with inlet and oven temperatures of 210" which may have caused some thermal decomposition. Each of the spectra contains several hundred peaks, so only the more abundant fragments or peaks corresponding to obvious fragments are reported. There were five fragments of large relative abundance appearing in both spectra: 15 (CH3), 30 (BF), 61 (CH&OF?), 73 (CHzCO2CH3), and 91 (CHzCOBFz?). The base peak for (FBNCHzC02CH3)3was a t m / e 49 (BFz?). The spectrum also contained a small peak a t 352, the parent ion, and a large peak a t 292 which is P - COZCH3. The loss of an ester group is the primary mode of fragmentation for a-amino esters.16 The base peak for the other borazine was a t m/e 31 (OCH3and/or HBF). The heaviest mass peak observed was 394 (P - F ? ) . A comparison of the ir spectra of CHzClZ solutions of F,B H2NCHzCHZCOzCH3and ( F B N C H Z C H Z C O ~ ) ~ shows no change in carbonyl stretching frequencies. This would seem to indicate the lack of any association between the carbonyl group and a ring boron. The IlB chemical shift of -35.7 ppm also is indicative of a fluoroborazine and not of a four-coordinate boron.

-

Acknowledgment.-We wish to acknowledge NSF support for nmr instrumentation. We wish to thank Dr. Kurt Niedenzu of the University of Kentucky for the mass spectra and Dr. Sheldon Shore of T h e Ohio State University for several of the boron nmr spectra. (15) H. S. Turner and R . J . Warne, Chem. Ind. (London), 526 (1958). (16) F. W. McLafferty, "Mass Spectra of Organic Ions," Academic Press, New York, N. Y . ,1961.